luke
/
sbt
mirror of https://github.com/sbt/sbt.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #4119 from dwijnand/cleanup/contributing 

Cleanup CONTRIBUTING.md
This commit is contained in:
eugene yokota 2018-04-25 13:22:21 -04:00 committed by GitHub
parent 564bbeab54 952858e68b
commit 72556b28c7
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
5 changed files with 194 additions and 192 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

209
CONTRIBUTING.md
View File

@ -1,33 +1,14 @@
[StackOverflow]: http://stackoverflow.com/tags/sbt
[ask]: https://stackoverflow.com/questions/ask?tags=sbt
[Setup]: http://www.scala-sbt.org/release/docs/Getting-Started/Setup
[Issues]: https://github.com/sbt/sbt/issues
[sbt-dev]: https://groups.google.com/d/forum/sbt-dev
[sbt-contrib]: https://gitter.im/sbt/sbt-contrib
[Lightbend]: https://www.lightbend.com/
[subscriptions]: https://www.lightbend.com/platform/subscription
[327]: https://github.com/sbt/sbt/issues/327
[gitter]: https://gitter.im/sbt/sbt
[documentation]: https://github.com/sbt/website
Support
=======
Contributing
============
[Lightbend] sponsors sbt and encourages contributions from the active community. Enterprises can adopt it for mission critical systems with confidence because Lightbend stands behind sbt with commercial support and services.
For community support please [ask] on StackOverflow with the tag "sbt".
- State the problem or question clearly and provide enough context. Code examples and `build.sbt` are often useful when appropriately edited.
- There's also [Gitter sbt/sbt room][gitter], but Stackoverflow is recommended so others can benefit from the answers.
For professional support, [Lightbend], the maintainer of Scala compiler and sbt, provides:
- [Lightbend Subscriptions][subscriptions], which includes Expert Support
- Training
- Consulting
How to contribute to sbt
========================
(For support, see [SUPPORT](./SUPPORT.md))
There are lots of ways to contribute to sbt ecosystem depending on your interests and skill level.
@ -47,9 +28,13 @@ When you find a bug in sbt we want to hear about it. Your bug reports play an im
Effective bug reports are more likely to be fixed. These guidelines explain how to write such reports and pull requests.
Please open a GitHub issue when you are 90% sure it's an actual bug.
If you have an enhancement idea, or a general discussion, bring it up to [sbt-contrib].
### Notes about Documentation
Documentation fixes and contributions are as much welcome as to patching the core. Visit [the website project][documentation] to learn about how to contribute.
Documentation fixes and contributions are as much welcome as to patching the core. Visit [sbt/website][documentation] to learn about how to contribute.
### Preliminaries
@ -58,35 +43,29 @@ Documentation fixes and contributions are as much welcome as to patching the cor
- Open one case for each problem.
- Proceed to the next steps for details.
### Where to get help and/or file a bug report
sbt project uses GitHub Issues as a publicly visible todo list. Please open a GitHub issue when you are 90% sure it's an actual bug.
- If you need help with sbt, please [ask] on StackOverflow with the tag "sbt" and the name of the sbt plugin if any.
- If you have an enhancement idea, or a general discussion, bring it up to [sbt-contrib].
- If you need a faster response time, consider one of the [Lightbend subscriptions][subscriptions].
### What to report
The developers need three things from you: **steps**, **problems**, and **expectations**.
### Steps
The most important thing to remember about bug reporting is to clearly distinguish facts and opinions.
The most important thing to remember about bug reporting is to clearly distinguish facts and opinions. What we need first is **the exact steps to reproduce your problems on our computers**. This is called *reproduction steps*, which is often shortened to "repro steps" or "steps." Describe your method of running sbt. Provide `build.sbt` that caused the problem and the version of sbt or Scala that was used. Provide sample Scala code if it's to do with incremental compilation. If possible, minimize the problem to reduce non-essential factors.
#### Steps
What we need first is **the exact steps to reproduce your problems on our computers**. This is called *reproduction steps*, which is often shortened to "repro steps" or "steps." Describe your method of running sbt. Provide `build.sbt` that caused the problem and the version of sbt or Scala that was used. Provide sample Scala code if it's to do with incremental compilation. If possible, minimize the problem to reduce non-essential factors.
Repro steps are the most important part of a bug report. If we cannot reproduce the problem in one way or the other, the problem can't be fixed. Telling us the error messages is not enough.
### Problems
#### Problems
Next, describe the problems, or what *you think* is the problem. It might be "obvious" to you that it's a problem, but it could actually be an intentional behavior for some backward compatibility etc. For compilation errors, include the stack trace. The more raw info the better.
### Expectations
#### Expectations
Same as the problems. Describe what *you think* should've happened.
### Notes
#### Notes
Add an optional notes section to describe your analysis.
Add any optional notes section to describe your analysis.
### Subject
@ -120,7 +99,7 @@ See below for the branch to work against.
### Adding notes
All pull requests are required to include a "Notes" file which documents the change. This file should reside in the
Most pull requests should include a "Notes" file which documents the change. This file should reside in the
directory:
<sbt root>
@ -234,164 +213,14 @@ command. To run a single test, such as the test in
Profiling sbt
-------------
There are several ways to profile sbt. The new hotness in profiling is FlameGraph.
You first collect stack trace samples, and then it is processed into svg graph.
See:
- [Using FlameGraphs To Illuminate The JVM by Nitsan Wakart](https://www.youtube.com/watch?v=ugRrFdda_JQ)
- [USENIX ATC '17: Visualizing Performance with Flame Graphs](https://www.youtube.com/watch?v=D53T1Ejig1Q)
### jvm-profiling-tools/async-profiler
The first one I recommend is async-profiler. This is available for macOS and Linux,
and works fairly well.
1. Download the installer from https://github.com/jvm-profiling-tools/async-profiler/releases/tag/v1.2
2. Make symbolic link to `build/` and `profiler.sh` to `$HOME/bin`, assuming you have PATH to `$HOME/bin`:
`ln -s ~/Applications/async-profiler/profiler.sh $HOME/bin/profiler.sh`
`ln -s ~/Applications/async-profiler/build $HOME/bin/build`
Next, close all Java appliations and anything that may affect the profiling, and run sbt in one terminal:
```
$ sbt exit
```
In another terminal, run:
```
$ jps
92746 sbt-launch.jar
92780 Jps
```
This tells you the process ID of sbt. In this case, it's 92746. While it's running, run
```
$ profiler.sh -d 60 <process id>
Started [cpu] profiling
--- Execution profile ---
Total samples: 31602
Non-Java: 3239 (10.25%)
GC active: 46 (0.15%)
Unknown (native): 14667 (46.41%)
Not walkable (native): 3 (0.01%)
Unknown (Java): 433 (1.37%)
Not walkable (Java): 8 (0.03%)
Thread exit: 1 (0.00%)
Deopt: 9 (0.03%)
Frame buffer usage: 55.658%
Total: 1932000000 (6.11%) samples: 1932
[ 0] java.lang.ClassLoader$NativeLibrary.load
[ 1] java.lang.ClassLoader.loadLibrary0
[ 2] java.lang.ClassLoader.loadLibrary
[ 3] java.lang.Runtime.loadLibrary0
[ 4] java.lang.System.loadLibrary
....
```
This should show a bunch of stacktraces that are useful.
To visualize this as a flamegraph, run:
```
$ profiler.sh -d 60 -f /tmp/flamegraph.svg <process id>
```
This should produce `/tmp/flamegraph.svg` at the end.
![flamegraph](project/flamegraph_svg.png)
See https://gist.github.com/eed3si9n/82d43acc95a002876d357bd8ad5f40d5
### running sbt with standby
One of the tricky things you come across while profiling is figuring out the process ID,
while wnating to profile the beginning of the application.
For this purpose, we've added `sbt.launcher.standby` JVM flag.
In the next version of sbt, you should be able to run:
```
$ sbt -J-Dsbt.launcher.standby=20s exit
```
This will count down for 20s before doing anything else.
### jvm-profiling-tools/perf-map-agent
If you want to try the mixed flamegraph, you can try perf-map-agent.
This uses `dtrace` on macOS and `perf` on Linux.
You first have to compile https://github.com/jvm-profiling-tools/perf-map-agent.
For macOS, here to how to export `JAVA_HOME` before running `cmake .`:
```
$ export JAVA_HOME=$(/usr/libexec/java_home)
$ cmake .
-- The C compiler identification is AppleClang 9.0.0.9000039
-- The CXX compiler identification is AppleClang 9.0.0.9000039
...
$ make
```
In addition, you have to git clone https://github.com/brendangregg/FlameGraph
In a fresh termimal, run sbt with `-XX:+PreserveFramePointer` flag:
```
$ sbt -J-Dsbt.launcher.standby=20s -J-XX:+PreserveFramePointer exit
```
In the terminal that you will run the perf-map:
```
$ cd quicktest/
$ export JAVA_HOME=$(/usr/libexec/java_home)
$ export FLAMEGRAPH_DIR=$HOME/work/FlameGraph
$ jps
94592 Jps
94549 sbt-launch.jar
$ $HOME/work/perf-map-agent/bin/dtrace-java-flames 94549
dtrace: system integrity protection is on, some features will not be available
dtrace: description 'profile-99 ' matched 2 probes
Flame graph SVG written to DTRACE_FLAME_OUTPUT='/Users/xxx/work/quicktest/flamegraph-94549.svg'.
```
This would produce better flamegraph in theory, but the output looks too messy for `sbt exit` case.
See https://gist.github.com/eed3si9n/b5856ff3d987655513380d1a551aa0df
This might be because it assumes that the operations are already JITed.
### ktoso/sbt-jmh
https://github.com/ktoso/sbt-jmh
Due to JIT warmup etc, benchmarking is difficult. JMH runs the same tests multiple times to
remove these effects and comes closer to measuring the performance of your code.
There's also an integration with jvm-profiling-tools/async-profiler, apparently.
### VisualVM
I'd also mention traditional JVM profiling tool. Since VisualVM is opensource,
I'll mention this one: https://visualvm.github.io/
1. First VisualVM.
2. Start sbt from a terminal.
3. You should see `xsbt.boot.Boot` under Local.
4. Open it, and select either sampler or profiler, and hit CPU button at the point when you want to start.
If you are familiar with YourKit, it also works similarly.
See [PROFILING](./PROFILING.md)
Other notes for maintainers
---------------------------
### Publishing VS Code Extensions
https://code.visualstudio.com/docs/extensions/publish-extension
Reference https://code.visualstudio.com/docs/extensions/publish-extension
```
$ sbt

2
ISSUE_TEMPLATE.md
View File

@ -1,4 +1,4 @@
(See the guidelines for contributing, linked above)
- [ ] I've read the [CONTRIBUTING](https://github.com/sbt/sbt/blob/1.x/CONTRIBUTING.md) guidelines
## steps

153
PROFILING.md Normal file
View File

@ -0,0 +1,153 @@
Profiling sbt
-------------
There are several ways to profile sbt. The new hotness in profiling is FlameGraph.
You first collect stack trace samples, and then it is processed into svg graph.
See:
- [Using FlameGraphs To Illuminate The JVM by Nitsan Wakart](https://www.youtube.com/watch?v=ugRrFdda_JQ)
- [USENIX ATC '17: Visualizing Performance with Flame Graphs](https://www.youtube.com/watch?v=D53T1Ejig1Q)
### jvm-profiling-tools/async-profiler
The first one I recommend is async-profiler. This is available for macOS and Linux,
and works fairly well.
1. Download the installer from https://github.com/jvm-profiling-tools/async-profiler/releases/tag/v1.2
2. Make symbolic link to `build/` and `profiler.sh` to `$HOME/bin`, assuming you have PATH to `$HOME/bin`:
`ln -s ~/Applications/async-profiler/profiler.sh $HOME/bin/profiler.sh`
`ln -s ~/Applications/async-profiler/build $HOME/bin/build`
Next, close all Java appliations and anything that may affect the profiling, and run sbt in one terminal:
```
$ sbt exit
```
In another terminal, run:
```
$ jps
92746 sbt-launch.jar
92780 Jps
```
This tells you the process ID of sbt. In this case, it's 92746. While it's running, run
```
$ profiler.sh -d 60 <process id>
Started [cpu] profiling
--- Execution profile ---
Total samples: 31602
Non-Java: 3239 (10.25%)
GC active: 46 (0.15%)
Unknown (native): 14667 (46.41%)
Not walkable (native): 3 (0.01%)
Unknown (Java): 433 (1.37%)
Not walkable (Java): 8 (0.03%)
Thread exit: 1 (0.00%)
Deopt: 9 (0.03%)
Frame buffer usage: 55.658%
Total: 1932000000 (6.11%) samples: 1932
[ 0] java.lang.ClassLoader$NativeLibrary.load
[ 1] java.lang.ClassLoader.loadLibrary0
[ 2] java.lang.ClassLoader.loadLibrary
[ 3] java.lang.Runtime.loadLibrary0
[ 4] java.lang.System.loadLibrary
....
```
This should show a bunch of stacktraces that are useful.
To visualize this as a flamegraph, run:
```
$ profiler.sh -d 60 -f /tmp/flamegraph.svg <process id>
```
This should produce `/tmp/flamegraph.svg` at the end.
![flamegraph](project/flamegraph_svg.png)
See https://gist.github.com/eed3si9n/82d43acc95a002876d357bd8ad5f40d5
### running sbt with standby
One of the tricky things you come across while profiling is figuring out the process ID,
while wnating to profile the beginning of the application.
For this purpose, we've added `sbt.launcher.standby` JVM flag.
In the next version of sbt, you should be able to run:
```
$ sbt -J-Dsbt.launcher.standby=20s exit
```
This will count down for 20s before doing anything else.
### jvm-profiling-tools/perf-map-agent
If you want to try the mixed flamegraph, you can try perf-map-agent.
This uses `dtrace` on macOS and `perf` on Linux.
You first have to compile https://github.com/jvm-profiling-tools/perf-map-agent.
For macOS, here to how to export `JAVA_HOME` before running `cmake .`:
```
$ export JAVA_HOME=$(/usr/libexec/java_home)
$ cmake .
-- The C compiler identification is AppleClang 9.0.0.9000039
-- The CXX compiler identification is AppleClang 9.0.0.9000039
...
$ make
```
In addition, you have to git clone https://github.com/brendangregg/FlameGraph
In a fresh termimal, run sbt with `-XX:+PreserveFramePointer` flag:
```
$ sbt -J-Dsbt.launcher.standby=20s -J-XX:+PreserveFramePointer exit
```
In the terminal that you will run the perf-map:
```
$ cd quicktest/
$ export JAVA_HOME=$(/usr/libexec/java_home)
$ export FLAMEGRAPH_DIR=$HOME/work/FlameGraph
$ jps
94592 Jps
94549 sbt-launch.jar
$ $HOME/work/perf-map-agent/bin/dtrace-java-flames 94549
dtrace: system integrity protection is on, some features will not be available
dtrace: description 'profile-99 ' matched 2 probes
Flame graph SVG written to DTRACE_FLAME_OUTPUT='/Users/xxx/work/quicktest/flamegraph-94549.svg'.
```
This would produce better flamegraph in theory, but the output looks too messy for `sbt exit` case.
See https://gist.github.com/eed3si9n/b5856ff3d987655513380d1a551aa0df
This might be because it assumes that the operations are already JITed.
### ktoso/sbt-jmh
https://github.com/ktoso/sbt-jmh
Due to JIT warmup etc, benchmarking is difficult. JMH runs the same tests multiple times to
remove these effects and comes closer to measuring the performance of your code.
There's also an integration with jvm-profiling-tools/async-profiler, apparently.
### VisualVM
I'd also mention traditional JVM profiling tool. Since VisualVM is opensource,
I'll mention this one: https://visualvm.github.io/
1. First VisualVM.
2. Start sbt from a terminal.
3. You should see `xsbt.boot.Boot` under Local.
4. Open it, and select either sampler or profiler, and hit CPU button at the point when you want to start.
If you are familiar with YourKit, it also works similarly.

2
PULL_REQUEST_TEMPLATE.md
View File

@ -1 +1 @@
(See the guidelines for contributing, linked above)
- [ ] I've read the [CONTRIBUTING](https://github.com/sbt/sbt/blob/1.x/CONTRIBUTING.md) guidelines

20
SUPPORT.md Normal file
View File

@ -0,0 +1,20 @@
[ask]: https://stackoverflow.com/questions/ask?tags=sbt
[Lightbend]: https://www.lightbend.com/
[subscriptions]: https://www.lightbend.com/platform/subscription
[gitter]: https://gitter.im/sbt/sbt
Support
=======
[Lightbend] sponsors sbt and encourages contributions from the active community. Enterprises can adopt it for mission critical systems with confidence because Lightbend stands behind sbt with commercial support and services.
For community support please [ask] on StackOverflow with the tag "sbt" (and the name of the sbt plugin(s) if any).
- State the problem or question clearly and provide enough context. Code examples and `build.sbt` are often useful when appropriately edited.
- There's also [Gitter sbt/sbt room][gitter], but Stackoverflow is recommended so others can benefit from the answers.
For professional support, for instance if you need faster response times, [Lightbend], the maintainer of Scala compiler and sbt, provides:
- [Lightbend Subscriptions][subscriptions], which includes Expert Support
- Training
- Consulting